【概要】　【スライド】

【論文】※4/14差替版

【論文】

要旨：Spatial data are characterized by their spatial dependence, which is often complex, non-linear, and difficult to fully capture with a single model. Significant levels of model uncertainty– arising from these characteristics– cannot be resolved by model selection or simple ensemble methods, as performances are not homogeneous. We address this issue by proposing a novel methodology that captures spatially-varying model uncertainty, which we call Bayesian spatial predictive synthesis. Our proposal is defined by specifying a latent factor spatially-varying coefficient model as the synthesis function, which enables model coefficients to vary over the region to achieve flexible spatial model ensembling. We show that our proposal is derived from the theoretically best approximation of the data generating process and that it provides a finite sample theoretical guarantee for its predictive performance, specifically that the predictions are exact minimax. Two MCMC strategies are implemented for full uncertainty quantification, as well as a variational inference strategy for fast point inference. We also extend the estimation strategy for general responses. Through simulation examples and two real data applications, we demonstrate that our proposed Bayesian spatial predictive synthesis outperforms standard spatial models and ensemble methods, and advanced machine learning methods, in terms of predictive accuracy, while maintaining interpretability of the prediction mechanism.

【論文】

要旨：Using panel data from 17 countries with varying economic circumstances from 1974 to 2019, we estimate regression models that explain residential property price dynamics by incorporating demographic factors and considering the interaction of those demographics with credit conditions. Our results show the importance of the demographic factors in modeling the long-run equilibrium of residential property prices. We find that the effect of nominal interest rates determined by monetary policy on asset prices varies depending on the country and the degree of population aging at the time. We also find that the persistently optimistic population projections lead to the oversupply of the residential stock in rapidly aging countries, resulting in stagnant residential property markets.

《Paper》

Abstract: We develop an urban growth model where human capital spillovers foster entrepreneurship and learning in heterogeneous cities. Incumbent residents limit city expansion through planning regulations so that commuting and housing costs do not outweigh productivity gains from agglomeration. The model builds on strong microfoundations, matches key regularities at the city and economy-wide levels, and generates novel predictions for which we provide evidence. It can be quantified relying on few parameters, provides a basis to estimate the main ones, and remains transparent regarding its mechanisms. We examine various counterfactuals to assess the effect of cities on economic growth and aggregate output quantitatively.

【論文１（PDF）】【論文２（PDF）】

要旨：New ideas and technologies adopted by a small number of individuals occasionally spread globally through a complex web of social ties. Here, we present a simple and general approximation method, namely, a message-passing approach, that allows us to describe the diffusion processes on (sparse) random networks in an almost exact manner. We consider two classes of binary-action games where the best pure strategies for individual players are characterized as variants of the threshold rule. We verify that the dynamics of diffusion observed on synthetic networks are accurately replicated by the message-passing equation, whose fixed point corresponds to a Nash equilibrium, while the conventional mean-field method tends to overestimate the size and frequency of diffusion. Generalized cascade conditions under which a global diffusion can occur are also provided. We extend the framework to analyze diffusion of multiple goods.

【論文（PDF）】

要旨：携帯電話位置情報データを用いると、国全体などの広範囲の長距離旅行量とその詳細な時間的な変化をかなり精度よく把握できる。このデータを用いて2015年の北陸新幹線・金沢開業前後で前後1年間の比較をすると、所要時間が短縮された場所の変化だけでなく、新幹線で接続されていない西日本各地から石川県への来訪者の増加も大きくみられることがわかる。本研究では、後者のような空間パターンの変化に着目し、その特徴を解明するために、多時点の居住地-旅行先分布表を分解して、近年に発生した類似の変化を探索的に検出することを試みた。具体的には、都道府県あるいは市区町村単位の居住地・旅行先ペアごとの旅行先選択確率の変化を、対称行列（交通サービス変化による直接的な変化を含む対称な変化）と旅行先ごとに均一な値の入る行列（全国から均一に旅行者数を増やす効果）に分解した。その結果として、(1)居住地-旅行先表の経年変化は２種類の空間パターンでほとんど説明できること、(2) 北陸新幹線開業では後者のパターン変化が大きかったこと、(3)後者の効果は3年以上継続しており短期的な広告効果ではないこと、(4)後者の効果がない新幹線開業地も存在すること、などを明らかにした。

要旨：With the rapid increase in the availability of passive data in the field of transportation, combining machine learning with transportation models has emerged as an important research topic in recent years. This study proposes an entropy Tucker model that integrates (1) a Tucker decomposition technique, i.e., an existing machine learning method, and (2) an entropy maximizing model, i.e., an existing model used for modeling trip distribution in the field of transportation. In addition, an optimization algorithm is presented to empirically identify the proposed model. The proposed model provides a solid theoretical foundation for the machine learning method, substantially improves prediction performance, and provides richer behavioral implications through empirical parameter estimation of travel impedance. We empirically apply the proposed method to evaluate the impacts of changes in the public transport fare structure on the destination choice of public transport users in Hiroshima by using the smart card data collected in Hiroshima, Japan. The estimated values of travel time range from 1.146 to 14.44 JPY/min, which is consistent with that reported in existing studies. The results of scenario analysis with different public transport fare structures suggest that identified changes in trip patterns, revenues, and users’ benefits for the public transport operator are considerably different between the conventional entropy model and the proposed entropy Tucker model. Further, we confirm that the users’ benefits vary based on the time of day. These obtained results confirm the importance of considering the heterogeneous preferences of users in economic appraisals.

要旨：土地利用モデルのパラメータは，実用上の問題により、いくつかの段階に分けて推定・キャリブレートされることが多い。しかし、理論的には一括で推定することが望ましく、ゆえに現状の推定手法の信頼性は定かでない。そこで、ベイズ法により全パラメータを一括で推定することを試みた。具体的には、素朴な土地利用モデルに対して、パラメータ推定、変数選択、現況再現性の確認、感度分析を行った。さらに、集積の経済の概念を導入したモデルへの拡張可能性も検討・検証した。推定結果や従来手法との比較から、ベイズ法を用いた推定手法の将来性が示唆された。